News Stay informed about the latest enterprise technology news and product updates.

Experts: Encryption not a security cure-all

Enterprises have to understand what security problems they have before deciding whether to encrypt data or secure the network or system on which that data is traveling.

Encryption is a lot like garlic. It's a versatile part of many security recipes, but not to the exclusion of other ingredients.

Some companies believe that, by encrypting data, they're solving most security problems, and they don't care whether attackers sniff packets or break into systems -- because the data will be useless to them if it's encrypted. But relying too much on encryption could lull companies into a false sense of security and keep them from paying necessary attention to network and operating system security, a tendency that could come back to haunt them.

In other words, encryption is not a solution to security concerns but a way to transfer the problem, said Benjamin Jun, vice president at Cryptography Research Inc., a San Francisco-based consulting company. "Instead of managing 100 GB of critical data, you only have to manage a key, which is much smaller in size," he said.

Before implementing any encryption, companies need to fully understand the security problem they want to solve. "This may sound easy, but it can be quite difficult," Jun said.

There are instances where encryption will solve particular concerns and other places where different solutions are needed. For example, protecting data on laptops is one problem many companies deal with. It's a fact of life that laptops will be lost or stolen. Encrypting the data on them makes a lot of sense.

On the other hand, Cryptography Research, in another example, was faced with a different problem. How could the company use e-mail and have standard Web access on the same network that it uses to transport sensitive customer data? Instead of turning to encryption, the company opted for two independent networks, one used for e-mail and Web browsing, and another used exclusively for sensitive data. "We couldn't proactively stop everything at the firewall," Jun said.

Does (key) size matter?

Companies have to consider many factors when considering encryption, like key size and the length of time they want to secure and store sensitive data.

Every few months, stories emerge about the latest key size to be cracked. In October, RSA Security's RC5-64 cipher, the company's 64-bit encryption key, was cracked. Now, to be fair, it took the equivalent of the population of a large city and its computing power nearly five years to break a single key. Using 64-bit encryption would be far less risky than other security activities.

Experts, however, warn against using 64-bit ciphers because of Moore's Law. In other words, computing power increases at such a rate that, in the not too distant future, systems would be able to crack such encryption in a much shorter amount of time.

So when one is considering a key size, the amount of time that data will need to be secure comes into play. In theory, encryption can be foiled by random tries, so a longer period of time gives an attacker a bigger window of opportunity to find the key.

Jun warns against focusing too much on key size. "It's like arguing about how good the deadbolt on your front door is. The lock may be great, but your house has windows which pose a much greater security risk," he said.

Much like a deadbolt-lover may leave more valuables on the kitchen table because of a false sense of security, people who rely too heavily on encryption may be opening themselves to other risks, Jun said. There have been some implementations that have been vulnerable because the random number generator was faulty, Jun said.

So what if you want to store some important data for 20 years? Well, Ronald L. Rivest and Adi Shamir, two of the developers of the RSA algorithm, recently suggested 2,048-bit keys for such long-term, important storage. They suggested this idea during a panel discussion on cryptography at the recent RSA Conference in San Francisco.

Jun recommends using 3 DES or AES (either 128-bit or 256-bit) for such long-term storage. But users shouldn't get too bogged down about the algorithm or key size they use. For the former, pick the ones that have been around for a while and are well tested. The more established encryption algorithms are better tested for security than virtually any other piece of the security infrastructure.

FOR MORE INFORMATION: news exclusive: "Expert deconstructs the cracking of 64-bit key"

Best Web Links on encryption


Dig Deeper on Disk and file encryption tools

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.